Development and validation of biomarkers related to anoikis in liver cirrhosis based on bioinformatics analysis

Abstract

BACKGROUND

According to study, anoikis-related genes (ARGs) have been demonstrated to play a significant impact in cirrhosis, a major disease threatening human health worldwide.

AIM

To investigate the relationship between ARGs and cirrhosis development to provide insights into the clinical treatment of cirrhosis.

METHODS

RNA-sequencing data related to cirrhosis were obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) between cirrhotic and normal tissues were intersected with ARGs to derive differentially expressed ARGs (DEARGs). The DEARGs were filtered using the least absolute shrinkage and selection operator, support vector machine recursive feature elimination, and random forest algorithms to identify biomarkers for cirrhosis. These biomarkers were used to create a nomogram for predicting the prognosis of cirrhosis. The proportions of diverse immune cell subsets in cirrhotic vs normal tissues were compared using the CIBERSORT computational method. In addition, the linkage between immune cells and biomarkers was assessed, and a regulatory network of mRNA, miRNA, and transcription factors was constructed relying on the biomarkers.

RESULTS

The comparison of cirrhotic and normal tissue samples led to the identification of 635 DEGs. Subsequent intersection of the DEGs with ARGs produced a set of 26 DEARGs. Subsequently, three DEARGs, namely, ACTG1, STAT1, and CCR7, were identified as biomarkers using three machine-learning algorithms. The proportions of M1 and M2 macrophages, resting CD4 memory T cells, resting mast cells, and plasma cells significantly differed between cirrhotic and normal tissue samples. The proportions of M1 and M2 macrophages, resting CD4 memory T cells, and resting mast cells were significantly correlated with the expression of the three biomarkers. The mRNA–miRNA–TF network showed that ACTG1, CCR7, and STAT1 were regulated by 28, 42, and 35 miRNAs, respectively. Moreover, AR, MAX, EP300, and FOXA1 were found to regulate four miRNAs related to the biomarkers.

CONCLUSION

This study revealed ACTG1, STAT1, and CCR7 as biomarkers of cirrhosis, providing a reference for developing novel diagnostic and therapeutic strategies for cirrhosis.

Key Words: Anoikis-related genes; Cirrhosis; Machine learning; Biomarker; Therapeutic drugs; Bioinformatics; Immune infiltration

Core Tip: Studies have highlighted the role of anoikis-related genes (ARGs) in cirrhosis. In this study, machine learning algorithms were used to identify differentially expressed ARGs (DEARGs) based on RNA-sequencing data. Three DEARGs were identified as biomarkers for cirrhosis (ACTG1, STAT1, and CCR7). The proportions of M1 and M2 macrophages, CD4 T cells, and mast cells were different between cirrhotic and normal tissues and were correlated with the expression of the three biomarkers. An mRNA–miRNA–TF network was constructed based on the three biomarkers. miRNAs and transcription factor regulating the biomarkers were identified. The findings of this study may facilitate the development of novel diagnostic and therapeutic strategies for cirrhosis.

INTRODUCTION

Cirrhosis marks the advanced, concluding phase of chronic liver disease. It^’s etiological factors vary geographically. Cirrhosis in western countries is predominantly caused by heavy alcohol consumption, hepatitis C, and non-alcoholic fatty liver disease. However, the leading driver of cirrhosis in the Asia-Pacific area is chronic hepatitis B. According to the Global Health Metrics findings data, in 2015, The age-adjusted prevalence of cirrhosis and chronic liver disease amounted to 20.7 per 100000 individuals, which was 13% higher than that reported in 2000[1]. A survey[2] reported that Roughly 1 million individuals succumb to complications of cirrhosis annually worldwide. The high prevalence of this liver disease worldwide imposes a significant strain on healthcare resources and societal well-being. The prognosis for those with end-stage cirrhosis is typically poor, as the disease has advanced to a critical point, with most patients progressing to decompensated cirrhosis and hepatocellular carcinoma. Moreover, in most cases, cirrhosis is diagnosed when the condition has reached an advanced stage. A recent study showed that liver fibrosis and early-stage cirrhosis are reversible, as liver fibrosis is a dynamic process. Accurate assessment and early intervention are key to delaying or reversing the progression of liver fibrosis and early-stage cirrhosis[3]. Therefore, identifying novel biomarkers and understanding their molecular mechanisms may facilitate the development of targeted treatments for early intervention, eventually improving the prognosis of cirrhosis.

Anoikis, a specialized form of programmed cell demise, transpires when cells become separated from the extracellular matrix. This phenomenon holds a pivotal function in organismal development, tissue equilibrium, disease manifestation, and neoplastic metastasis[4]. It is a specialized form of programmed cell death, is a subtype of apoptosis in which cellular demise transpires through the conventional apoptotic cascade[5]. Moreover, it is implicated in the pathogenesis of a range of cancers, such as ovarian carcinoma, salivary gland adenoid cystic carcinoma, lung adenocarcinoma, and hepatocellular carcinoma[6-9]. In patients with cirrhosis, hepatic progenitor/stem cells and cytotoxic CD8+ T cells undergo apoptosis, which may accelerate cancer development. A study[10] demonstrated that hepatocyte growth factor signaling is crucial for the promotion of normal and damaged liver CD8+ T cells. In addition, the c-Met receptor is involved in the FAS-induced cell apoptosis pathway. Necrosis is another type of cell death significantly associated with programmed hepatocyte death[11]. The selective triggering of diverse programmed cell death mechanisms not only affects the course of liver diseases but also offers new possibilities for therapeutic approaches. Anoikis has been associated with hepatocellular carcinoma and resistance to chemotherapeutic drugs[12]. While liver cirrhosis is a major predisposing condition for hepatocellular carcinoma, the existing literature on the association between liver cirrhosis and anoikis, a specific mode of cell death, is limited.

The primary objective of this research was to analyze the association between genes involved in anoikis, a form of programmed cell death, and the development of liver cirrhosis. RNA-sequencing data were extracted from the Gene Expression Omnibus database, and several bioinformatic techniques were used to identify cirrhosis biomarkers, analyze their functions, and explore their mechanisms in the development of cirrhosis from a genetic perspective. The conclusions drawn from this study may yield important implications for the therapeutic approach to cirrhosis.

MATERIALS AND METHODS

Collection of data

RNA-sequencing data related to cirrhosis were extracted from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/). The GSE89377 dataset including 12 cirrhotic and 13 normal liver tissue samples was used as the training cohort, whereas the GSE14323 dataset including 40 cirrhotic and 19 normal liver tissue samples was used as the validation cohort. In addition, 549 ARGs (relevance score > 7) were identified using the GeneCards database (https://www.genecards.org/).

Differential expression analysis between cirrhotic and normal tissue samples

The “limma” (version 3.52.2) package was used to detect differentially expressed genes (DEGs) between cirrhotic and normal liver tissue samples in the training dataset, with the screening threshold set to an adjusted P value of < 0.05 and a |log2FoldChange| value of > 0.5[13]. The identified DEGs were then intersected with a set of antibiotic resistance genes (ARGs) to obtain a subset of differentially expressed ARGs (DEARGs). Subsequently, the DEARGs were subjected to enrichment analysis based on the gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. An adjusted P value of < 0.05 indicated significant enrichment.

Screening of cirrhosis biomarkers

DEARGs were individually filtered using the least absolute shrinkage and selection operator (LASSO), support vector machine recursive feature elimination (SVM-RFE), and random forest (RF) algorithms. The feature genes identified using these machine learning algorithms were intersected to obtain biomarkers. To evaluate the ability of the identified biomarkers to differentiate between cirrhotic and normal tissue samples, a principal component analysis was conducted. Receiver operating characteristic curves were plotted to evaluate the predictive performance of the biomarkers in both training and validation datasets. The expression of the biomarkers was compared between cirrhotic and normal tissue samples, and a nomogram incorporating the biomarkers was generated. Calibration curves were plotted and decision curve analysis was performed to evaluate the predictive performance of the nomogram.

Enrichment and immune infiltration analyses of biomarkers

Gene set enrichment analysis (GSEA) was conducted to assess the enrichment of each biomarker in GO terms and KEGG pathways, with the screening threshold set to P values of < 0.05 and |Normalized Enrichment Score| values of > 1. Ingenuity pathway analysis was leveraged to pinpoint signaling pathways that exhibited statistically significant modulation by biomarker regulation. The CIBERSORT algorithm was used to evaluate the proportions of various immune cell types in cirrhotic and normal tissue samples. Subsequently, the correlation between differential immune cells and the three biomarkers was analyzed.

Construction of an mRNA–miRNA–TF network

miRNAs regulating the biomarkers were predicted using the miRWalk (http://mirwalk.umm.uni-heidelberg.de/) and Starbase (https://starbase.sysu.edu.cn/index.php) databases. The two miRNA sets were intersected to obtain overlapping miRNAs, and an mRNA–miRNA network was subsequently generated. The TransmiR database was used to predict transcription factors (TFs) based on the overlapping miRNAs, followed by the construction of a miRNA–TF network. Finally, the two aforementioned networks were merged to obtain an mRNA–miRNA–TF network. In addition, potential drugs targeting the three biomarkers were predicted using the DGIdb database (https://dgidb.genome.wustl.edu/). These drugs were screened and visualized in a network using the PubChem database (https://pubchem.ncbi.nlm.nih.gov/).

Reverse transcription–quantitative polymerase chain reaction

In total, 10 cirrhotic and 10 normal blood samples were obtained from the patients who provided consent from The Third People’s Hospital of Yunnan Province. This study was approved by the ethics committee of Third People’s Hospital of Yunnan Province. and all participants signed an informed consent form. All samples were lysed to extract total RNA. The extracted RNA was reverse-transcribed using the SureScript First-Strand cDNA Synthesis Kit (Servicebio, China). The cDNA samples were analyzed via quantitative real-time PCR employing the Universal Blue SYBR Green qPCR Master Mix (Servicebio, China) on the CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, United States). Uantification of the three biomarker RNA transcripts was performed, with glyceraldehyde-3-phosphate dehydrogenase serving as the endogenous control. The primer sequence of the biomarker is shown in Table 1.

Table 1 Primer sequences.

Biomarkers	Sequences
ACTG1 F	CCAGCTCTCGCACTCTGTT
ACTG1 R	ATTGCGACCCCGCCTTTTG
STAT1 F	GCACCAGAGCCAATGGAACTT
STAT1 R	GAGCCCACTATCCGAGACACC
CCR7 F	TCACGGACGATTACATCGGAG
CCR7 R	ATGATAGGGAGGAACCAGGCT
GAPDH F	CGAAGGTGGAGTCAACGGATTT
GAPDH R	ATGGGTGGAATCATATTGGAAC

RESULTS

DEARGs obtained via differential expression analysis

A transcriptomic assessment contrasting cirrhotic and normal tissue samples uncovered 635 genes with altered expression levels (Figure 1A and B). The intersection of these DEGs with ARGs revealed 26 DEARGs (Figure 1C). Functional enrichment analysis demonstrated that the DFG were overrepresented in 841 GO annotations, including nephron development, postsynaptic actin cytoskeleton, and postsynaptic actin cytoskeleton (Figure 1D), and 37 KEGG pathways, including focal adhesion, fluid shear stress and atherosclerosis, and regulation of actin cytoskeleton (Figure 1E).

Figure 1 Identification and functional enrichment analysis of differentially expressed anoikis-related genes. A: Volcano maps of differentially expressed genes (DEGs); B: Heat map of DEGs; C: Venn diagram of differentially expressed anoikis-related genes (DEARGs); D: Gene ontology analysis of DEARGs based on three domains, namely, cellular component, molecular function, and biological process; E: Kyoto Encyclopedia of Genes and Genome pathway analysis. DEGs: Differentially expressed gene; ARG: Anoikis-related gene.

ACTG1, STAT1, and CCR7 as biomarkers with superior diagnostic efficacy for liver cirrhosis

The 26 DEARGs were individually filtered using three machine-learning algorithms. The LASSO algorithm revealed 6 DEARGs, namely, ACTG1, STAT1, RAC1, CCR7, SOX9, and CCND1, as feature genes (Figure 2A and B). Similarly, the SVM-RFE algorithm revealed 3 feature genes, namely, ACTG1, STAT1, and CCR7; and the RF algorithm revealed 10 feature genes, namely, SOX9, CD24, RAC1, ACTG1, STAT1, CCR7, VIM, ACTB, PDGFRA, and SORT1 (Figure 2C and D). The three gene sets were intersected, resulting in the identification of three biomarkers for cirrhosis, namely, ACTG1, STAT1, and CCR7 (Figure 2E). Principal component analysis showed that the biomarkers effectively distinguished between cirrhotic and normal liver tissue samples (Figure 3A). In addition, the AUC values of the biomarkers were > 0.75 in both training and validation datasets (Figure 3B and C). Cirrhotic tissue samples displayed substantially increased biomarker expression levels in contrast to normal tissue samples (Figure 3D). Subsequently, a nomogram incorporating the A panel of three biomarkers was developed to forecast the prognosis for cirrhosis patients (Figure 3E). The C-index value of the calibration curve was 1, and decision curve analysis showed that all biomarkers had good predictive performance (Figure 3F and G).

Figure 2 Identification of biomarkers through machine learning algorithms. A: Least absolute shrinkage and selection operator coefficients profiles; B: Cross-validation for tuning parameter selection. Variables with non-zero coefficients were excluded; C: When n was 3, the classifier had the minimum error in the support vector machine recursive feature elimination model; D: Interpretation chart of two important evaluation indices; E: Venn diagram showing three biomarkers for cirrhosis.

Figure 3 Assessment and validation of the diagnostic efficacy of biomarkers. A: Principal component analysis was performed to analyze whether the three biomarkers could distinguish between control and cirrhotic samples; B: Receiver operating characteristic (ROC) curves of the biomarkers in the GSE89377 dataset; C: ROC curves of the biomarkers in the GSE14323 dataset; D: Expression of the biomarkers in the GSE89377 and GSE14323 datasets; E: A nomogram based on the biomarkers; F: Calibration curve of the nomogram; G: Decision curve analysis of the nomogram. AUC: Area under curve.

Significant negative correlation between M2 macrophages and the biomarkers

GSEA showed that ACTG1 was enriched in 976 GO terms, including cytoplasmic translation and cytosolic ribosome (Figure 4A), and 95 KEGG pathways, including focal adhesion and ribosome (Figure 4B). CCR7 was enriched in 903 GO terms, including cytoplasmic translation and ribosomal subunit (Figure 4C), and 78 KEGG pathways, including allograft rejection and intestinal immune network for IgA production (Figure 4D). STAT1 was enriched in 1056 GO terms, including MHC protein complex and antigen processing and presentation (Figure 4E), and 78 KEGG pathways, including allograft rejection and intestinal immune network for IgA production (Figure 4F). In addition, the biomarkers were found to play a role in 30 classical pathways; among which, the idiopathic pulmonary fibrosis-related signaling pathway with the highest Z-score was selected for subsequent analysis (Figure 5A and B). The CIBERSORT algorithm was employed to quantify the fractional representation of various immune cell subsets (Figure 5C). Proportional assessment revealed the representation of M1 macrophages, M2 macrophages, resting CD4 memory T cells, resting mast cells, and plasma cells significantly differed between cirrhotic and normal tissue samples (Figure 5D). Analysis revealed a significant negative relationship between the biomarker expression levels and the M2 macrophage proportion. In contrast, the biomarker expression levels were found to have a significant positive association with the proportions of resting CD4 memory T cells, resting mast cells, and M1 macrophages (Figure 5E).

Figure 4 Gene set enrichment analysis. A: Gene ontology (GO) enrichment analysis of ACTG1; B: Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of ACTG1; C: GO enrichment analysis of CCR7; D: KEGG enrichment analysis of CCR7; E: GO enrichment analysis of STAT1; F: KEGG enrichment analysis of STAT1. COVID-19: Coronavirus disease 2019.

Figure 5 Analysis of classical signaling pathways and immune infiltration. A: Classical pathways involving the three biomarkers; B: Idiopathic pulmonary fibrosis-related signaling pathway; C: Proportional stacking plot of immune cell distribution in each sample; D: Boxplot of immune cell distribution between the cirrhosis and control groups; E: Correlation between the biomarkers and differential immune cells.

Hsa-mir-320b as a regulator of the three biomarkers

A total of 62 miRNAs targeting the biomarkers were predicted using data from the miRWalk and Starbase databases. TFs were predicted based on these miRNAs, and an mRNA–miRNA–TF regulatory network was eventually constructed (Figure 6A–C). According to this network, the biomarkers were primarily regulated by seven miRNAs (hsa-mir-320b, hsa-mir-1294, hsa-mir-4525, hsa-mir-4429, hsa-mir-378b, hsa-mir-378a, and hsa-mir-520h). In particular, hsa-mir-320b was found to regulate all three biomarkers (Figure 6D). Furthermore, six potential drugs targeting STAT1 were identified (picoplatin, cisplatin, garcinol, guttiferone K, ipriflavone, and chembl85826), whereas only one potential drug targeting ACTG1 was identified (vincristine) (Figure 6E).

Figure 6 Analysis of classical signaling pathways and immune infiltration. A: Classical pathways involving the three biomarkers; B: Idiopathic pulmonary fibrosis-related signaling pathway; C: Proportional stacking plot of immune cell distribution in each sample; D: Boxplot of immune cell distribution between the cirrhosis and control groups; E: Correlation between the biomarkers and differential immune cells.

Significant upregulation of STAT1 and CCR7 in clinical cirrhotic tissue samples

Finally, the expression of the three biomarkers was assessed using quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) in a set of clinical specimens. The results provided evidence of the expression of STAT1 and CCR7 was significantly higher in cirrhotic tissue samples. The expression patterns of all three biomarkers in clinical samples were consistent with those observed via bioinformatic analysis of RNA-seq data (Figure 7).

Figure 7 Expression of biomarkers in clinical samples. A: ACTG1; B: STAT1; C: CCR7; NS: Not significant.

DISCUSSION

Anoikis is a form of programmed cell death that occurs through the canonical apoptotic mechanism[5]. Prior studies[6-9] have revealed anoikis' participation in the pathogenesis of an array of cancerous conditions, such as ovarian cancer, salivary gland adenoid cystic carcinoma, lung adenocarcinoma, and hepatocellular carcinoma. Although the relationship between anoikis and hepatocellular carcinoma is well-established, the relationship between cirrhosis, an important risk factor for hepatocellular carcinoma, and anoikis remains unclear. In this study, we identified three ARGs as biomarkers for cirrhosis via bioinformatic analysis. The expression of these biomarkers was found to be significantly associated with the abundance of M1 and M2 macrophages, resting CD4 memory T cells, and resting mast cells. Altogether, the findings of this study suggest that ARGs play an important role in the development of cirrhosis.

In this study, 26 DEARGs were identified between cirrhotic and normal liver tissue samples in the GSE89377 dataset. Functional enrichment analysis revealed the DEARGs to be strongly linked to the organization and dynamics of the postsynaptic actin cytoskeleton, as well as focal adhesion formation and actin cytoskeleton regulation. These signaling pathways are associated with cytoskeletal remodeling and extracellular matrix, playing diverse roles. Cell–extracellular matrix adhesion plays an essential role in various biological processes, including cell migration, differentiation, proliferation, and survival and gene regulation. Focal adhesions are specialized junctions that form at the interface between a cell and its extracellular matrix. At these sites, bundles of actin filaments within the cell are physically linked to transmembrane integrin receptors through intricate multi-protein complexes. Focal adhesions serve as structural hubs that integrate signaling pathways with the actin cytoskeleton and cell membrane receptors[13-16]. Focal adhesion kinase (FAK) plays a pivotal role in the activation and differentiation of hepatic stellate cells, a key driver of liver fibrosis. In fibrotic liver tissues, FAK activation correlates with increased expression of alpha-smooth muscle actin and collagen. The pro-fibrotic cytokine transforming growth factor beta-1 induces FAK activation in a dose- and time-dependent manner. Inhibiting FAK activation can downregulate the expression of collagen and alpha-smooth muscle actin, as well as reduce the formation of stress fibers[16,17]. An animal study (2020) showed that liver epithelial cells lacking FAK exhibited more severe injury and fibrosis than control cells[17-19]. Furthermore, a study on proteomic and transcriptomic data revealed a significant correlation between focal adhesion and the actin cytoskeleton in rats with liver fibrosis[18,19]. These findings are consistent with those of the present study.

Three machine learning algorithms were used to screen the 26 DEARGs to identify biomarkers for cirrhosis. The results revealed three DEARGs as good diagnostic indicators of cirrhosis, namely ACTG1, STAT1, and CCR7. The expression profiles of these biomarkers were consistent across the training and validation datasets. ACTG1, also known as γ-actin, is one of the six operational actin isoforms; among which, four isoforms are associated with muscle contraction, whereas the two remaining isoforms (ACTG1 and ACTB) serve essential functions in maintaining the structure of the cytoskeleton[19-21]. Existing evidence indicates that ACTG1 is implicated in the pathogenesis of various medical conditions. For instance, RRAD exhibits anti-tumor activity in hepatocellular carcinoma by reducing glucose metabolism and ACTG1 levels, thereby suppressing cell growth and division and promoting programmed cell death. These findings suggest that ACTG1 functions as a downstream mediator of RRAD, serving as a promising therapeutic target for hepatocellular carcinoma[21,22]. ACTG1 regulates cancer cell proliferation, migration, and death[22-24]; however, its role in cirrhosis remains unclear. STAT1, a dormant cytoplasmic TF, becomes activated in response to type I, II, and III interferons as well as interleukin-27. This protein plays an crucial role in the regulation of the Janus kinase (JAK)-STAT signaling pathway, thereby modulating immune responses[24-26]. This reduced activation of STAT1 in the JAK-STAT pathway appears to be a contributing factor in the decreased responsiveness to interferon-alpha observed during hepatitis B virus infection[26,27]. Dendritic cells and T lymphocytes, two critical cell types involved in adaptive immunity, both exhibit expression of the CCR7 receptor. It plays a key role in directing these cells to lymphoid organs. It is involved in immune tolerance and responses and is correlated with the prognosis of different types of tumors[27-29].

GSEA showed that the three biomarkers were significantly enriched in cytoplasmic translation, cytoplasmic ribosomes, allograft rejection, MHC protein complex, antigen processing and presentation, and intestinal immune network for IgA production. These pathways are associated with translation and immune responses. The biomarkers may promote the development of cirrhosis by activating these pathways. However, at present, the role of these pathways in liver fibrosis remains unknown. The findings of this study provide a valuable basis for investigating the involvement of these pathways in liver cirrhosis.

Immune infiltration analysis showed that the proportions of M1 and M2 macrophages, resting mast cells, plasma cells, and resting CD4 memory T cells were significantly higher in the cirrhosis group than in the control group. The expression levels of certain biomarkers exhibited a significant inverse correlation with the proportion of M2 macrophages. In contrast, these biomarkers were found to be significantly positively associated with the prevalence of resting CD4 memory T cells, resting mast cells, and M1 macrophages. Macrophages play an important role in innate immunity and exhibit significant heterogeneity by undergoing polarization. In response to certain microenvironmental changes in vivo and in vitro, such as cytokine secretion, macrophages can differentiate into M1 and M2 phenotypes, exhibiting functional differences. This phenomenon is called polarization[39,30]. The JAK/STAT1 and TLR4/NF-κB signaling pathways are important for M1 polarization, whereas STAT6 is primarily associated with M2 polarization[30-33]. Xie et al[34] showed that proteins interacting with C kinase 1 promoted the expression of STAT6 and p38α while inhibiting the NF-κB signaling pathway both in vivo and in vitro. These changes promoted M2 polarization and inhibited M1 polarization, thereby alleviating liver damage. Beljaars et al[35] conducted a study to localize and quantify the different phenotypes of macrophages in the liver tissues of patients with liver fibrosis. The results indicated that the proportions of both M1 and M2 macrophages were significantly elevated during the early stages of fibrosis[35,36]. Compared to M2 macrophages, M1 macrophages seem to have a more crucial part in the reversal of liver fibrosis[34,35]. M1 macrophages exacerbate liver damage via pro-inflammatory responses, whereas M2 macrophages alleviate liver injury through their anti-inflammatory and tissue-repairing effects[35,36]. Some bioinformatic studies[36,37] have shown that the expression of resting CD4+ memory T cells is monumentally higher in patients with non-alcoholic steatohepatitis than in healthy individuals. Similarly, some research have indicated a paramount connection between primary biliary cirrhosis and resting CD4+ memory T cells[37,38]. In addition, the clearance of CD4+ T cells has been displayed to decline liver swelling and fibrosis in mice.

On predicting drugs targeting the three biomarkers, we found that vincristine could target ACTG1, whereas picoplatin, cisplatin, garcinol, guttiferone K, ipriflavone, and chembl85826 could target STAT1. Vincristine is a chemotherapeutic drug commonly used to treat various malignant tumors, such as acute lymphoblastic leukemia and lymphoma. According to previous case reports, vincristine may exacerbate liver damage and increase the risk of septic shock and necrotizing fasciitis, particularly infections caused by Pseudomonas aeruginosa, in patients with blood system diseases complicated by hepatitis, cirrhosis, and other potential chronic liver diseases[38,39]. Some studies[39,40] have shown that hepatic arterial infusion chemotherapy with cisplatin can lead to hepatotoxic reactions, thereby inducing fibrosis, in patients with cirrhosis and advanced-stage hepatocellular carcinoma. Although existing studies have several limitations, both vincristine and cisplatin can exacerbate liver damage, as they have a similar structure. To assess the potential of vincristine and cisplatin in delaying the progression of liver fibrosis and cirrhosis, future studies should consider the use of cell and animal models of gene knockout. We speculate that ipriflavone, garcinol, guttiferone k, chembl85826, and picoplatin are promising drugs for treating cirrhosis. Further experiments and clinical trials should be performed to evaluate the clinical utility of these drugs.

The relative expression levels of the three biomarkers were validated in healthy and cirrhotic blood samples using qRT-PCR. The study revealed that the differential expression of ACTG1, STAT1, and CCR7 was significantly upregulated in the cirrhosis tissue samples, which was consistent with the results of bioinformatic analysis of transcriptomic data. Although the expression of ACTG1 was higher in cirrhotic tissue samples than in normal tissue samples, the difference was not statistically significant. This phenomenon may be attributed to variations in transcriptomic sequencing reads. Moreover, some inconsistencies may exist between qRT-PCR and transcriptomic sequencing data owing to differences in the detection methods.

CONCLUSION

The three biomarkers identified via bioinformatic analysis exhibit good diagnostic efficacy for cirrhosis. The results of enrichment analysis, immune infiltration analysis, and drug prediction provide novel venues for elucidating the mechanisms of the biomarkers and evaluating their use as therapeutic drugs for cirrhosis in clinical settings.